Troubleshooting Problems During a Failover Operation

Although it is possible for a failover operation to fail, it is unlikely. If an error does occur there are guidelines you can follow to fix the problem and then retry the broker failover.

Failed Failovers to Physical Standby Databases

These steps describe how to recover from a failed broker failover to a physical standby database.

Failed Broker Complete Physical Failovers

Examine the alert log file and the broker log file (drc*.log) on the target standby database to determine the cause of the failure and correct the problem.

If the reported problem can be corrected, then retry the failover operation. If the reported problem cannot be corrected or if the failover operation fails again after the reported problem has been corrected, then take the following steps:

  1. Connect to the target standby database and disable fast-start failover using the FORCE option if it is enabled.

  2. Then you can either:

    • Connect to another physical standby database and attempt a broker complete failover.

    • Perform a broker immediate failover to the target physical standby database.

  3. Reinstate the original primary database and any bystander physical standby databases that are disabled with a status of reinstatement required (ORA-16661).

  4. Reenable fast-start failover if it was disabled in step 1.

Failed Broker Immediate Physical Failovers

Examine the alert log file and the broker log file (drc*.log) on the target standby database to determine the cause of the failure and correct the problem.

If the problem can be corrected, retry the broker immediate failover. Otherwise connect to another physical standby database and attempt either a broker complete or immediate failover.

Failed Failovers to Logical Standby Databases

These are the steps to follow if a failover to a logical standby database fails.

  1. Examine the alert log file and the broker log file (drc*.log) on the target standby database to determine the cause of the failure and correct the problem.

  2. Connect to the target standby database and disable fast-start failover using the FORCE option if it is enabled.

  3. Retry the broker failover.

  4. Reinstate the old primary database. All bystander standby databases will be re-created from a copy of the new primary database.

  5. Reenable fast-start failover if it was disabled in step 1.

If broker failover continues to fail, you should stop the broker on all databases in the Oracle Data Guard configuration (set the DG_BROKER_START initialization parameter to FALSE). Remove the Oracle Data Guard broker configuration files from all databases. Attempt a manual failover using the guidelines for role transitions in Oracle Data Guard Concepts and Administration.

Note:

You can enable or disable the broker configuration using DGMGRL ENABLE CONFIGURATION and DISABLE CONFIGURATION commands. You cannot disable the configuration using Cloud Control. You can only enable the configuration using Cloud Control if it was previously disabled using DGMGRL.